osd_client.c 66.0 KB
Newer Older
1

2
#include <linux/ceph/ceph_debug.h>
S
Sage Weil 已提交
3

4
#include <linux/module.h>
S
Sage Weil 已提交
5 6 7 8 9 10
#include <linux/err.h>
#include <linux/highmem.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/slab.h>
#include <linux/uaccess.h>
11 12 13
#ifdef CONFIG_BLOCK
#include <linux/bio.h>
#endif
S
Sage Weil 已提交
14

15 16 17 18 19 20
#include <linux/ceph/libceph.h>
#include <linux/ceph/osd_client.h>
#include <linux/ceph/messenger.h>
#include <linux/ceph/decode.h>
#include <linux/ceph/auth.h>
#include <linux/ceph/pagelist.h>
S
Sage Weil 已提交
21

S
Sage Weil 已提交
22 23
#define OSD_OP_FRONT_LEN	4096
#define OSD_OPREPLY_FRONT_LEN	512
24

25
static const struct ceph_connection_operations osd_con_ops;
S
Sage Weil 已提交
26

A
Alex Elder 已提交
27
static void __send_queued(struct ceph_osd_client *osdc);
28
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd);
29 30 31 32
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req);
static void __unregister_linger_request(struct ceph_osd_client *osdc,
					struct ceph_osd_request *req);
33 34
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req);
S
Sage Weil 已提交
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

/*
 * Implement client access to distributed object storage cluster.
 *
 * All data objects are stored within a cluster/cloud of OSDs, or
 * "object storage devices."  (Note that Ceph OSDs have _nothing_ to
 * do with the T10 OSD extensions to SCSI.)  Ceph OSDs are simply
 * remote daemons serving up and coordinating consistent and safe
 * access to storage.
 *
 * Cluster membership and the mapping of data objects onto storage devices
 * are described by the osd map.
 *
 * We keep track of pending OSD requests (read, write), resubmit
 * requests to different OSDs when the cluster topology/data layout
 * change, or retry the affected requests when the communications
 * channel with an OSD is reset.
 */

/*
 * calculate the mapping of a file extent onto an object, and fill out the
 * request accordingly.  shorten extent as necessary if it crosses an
 * object boundary.
 *
 * fill osd op in request message.
 */
61
static int calc_layout(struct ceph_file_layout *layout, u64 off, u64 *plen,
62
			u64 *objnum, u64 *objoff, u64 *objlen)
S
Sage Weil 已提交
63
{
64
	u64 orig_len = *plen;
S
Sage Weil 已提交
65
	int r;
S
Sage Weil 已提交
66

67
	/* object extent? */
68 69
	r = ceph_calc_file_object_mapping(layout, off, orig_len, objnum,
					  objoff, objlen);
S
Sage Weil 已提交
70 71
	if (r < 0)
		return r;
72 73
	if (*objlen < orig_len) {
		*plen = *objlen;
74 75 76 77
		dout(" skipping last %llu, final file extent %llu~%llu\n",
		     orig_len - *plen, off, *plen);
	}

78
	dout("calc_layout objnum=%llx %llu~%llu\n", *objnum, *objoff, *objlen);
S
Sage Weil 已提交
79

80
	return 0;
S
Sage Weil 已提交
81 82
}

A
Alex Elder 已提交
83 84 85 86 87 88
static void ceph_osd_data_init(struct ceph_osd_data *osd_data)
{
	memset(osd_data, 0, sizeof (*osd_data));
	osd_data->type = CEPH_OSD_DATA_TYPE_NONE;
}

89
static void ceph_osd_data_pages_init(struct ceph_osd_data *osd_data,
90 91 92 93 94 95 96 97 98 99 100
			struct page **pages, u64 length, u32 alignment,
			bool pages_from_pool, bool own_pages)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGES;
	osd_data->pages = pages;
	osd_data->length = length;
	osd_data->alignment = alignment;
	osd_data->pages_from_pool = pages_from_pool;
	osd_data->own_pages = own_pages;
}

101
static void ceph_osd_data_pagelist_init(struct ceph_osd_data *osd_data,
102 103 104 105 106 107 108
			struct ceph_pagelist *pagelist)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_PAGELIST;
	osd_data->pagelist = pagelist;
}

#ifdef CONFIG_BLOCK
109
static void ceph_osd_data_bio_init(struct ceph_osd_data *osd_data,
110 111 112 113 114 115 116 117
			struct bio *bio, size_t bio_length)
{
	osd_data->type = CEPH_OSD_DATA_TYPE_BIO;
	osd_data->bio = bio;
	osd_data->bio_length = bio_length;
}
#endif /* CONFIG_BLOCK */

118 119 120 121 122 123
#define osd_req_op_data(oreq, whch, typ, fld)	\
	({						\
		BUG_ON(whch >= (oreq)->r_num_ops);	\
		&(oreq)->r_ops[whch].typ.fld;		\
	})

124 125
struct ceph_osd_data *
osd_req_op_extent_osd_data(struct ceph_osd_request *osd_req,
126
			unsigned int which)
127
{
128
	return osd_req_op_data(osd_req, which, extent, osd_data);
129 130 131 132 133 134 135
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data);

struct ceph_osd_data *
osd_req_op_cls_response_data(struct ceph_osd_request *osd_req,
			unsigned int which)
{
136
	return osd_req_op_data(osd_req, which, cls, response_data);
137 138 139 140
}
EXPORT_SYMBOL(osd_req_op_cls_response_data);	/* ??? */

void osd_req_op_extent_osd_data_pages(struct ceph_osd_request *osd_req,
141 142
			unsigned int which, struct page **pages,
			u64 length, u32 alignment,
143 144 145 146
			bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

147
	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
148 149 150 151 152 153
	ceph_osd_data_pages_init(osd_data, pages, length, alignment,
				pages_from_pool, own_pages);
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data_pages);

void osd_req_op_extent_osd_data_pagelist(struct ceph_osd_request *osd_req,
154
			unsigned int which, struct ceph_pagelist *pagelist)
155 156 157
{
	struct ceph_osd_data *osd_data;

158
	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
159 160 161 162 163 164
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data_pagelist);

#ifdef CONFIG_BLOCK
void osd_req_op_extent_osd_data_bio(struct ceph_osd_request *osd_req,
165
			unsigned int which, struct bio *bio, size_t bio_length)
166 167
{
	struct ceph_osd_data *osd_data;
168 169

	osd_data = osd_req_op_data(osd_req, which, extent, osd_data);
170 171 172 173 174 175 176 177 178 179 180
	ceph_osd_data_bio_init(osd_data, bio, bio_length);
}
EXPORT_SYMBOL(osd_req_op_extent_osd_data_bio);
#endif /* CONFIG_BLOCK */

static void osd_req_op_cls_request_info_pagelist(
			struct ceph_osd_request *osd_req,
			unsigned int which, struct ceph_pagelist *pagelist)
{
	struct ceph_osd_data *osd_data;

181
	osd_data = osd_req_op_data(osd_req, which, cls, request_info);
182 183 184
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}

185 186 187 188 189 190
void osd_req_op_cls_request_data_pagelist(
			struct ceph_osd_request *osd_req,
			unsigned int which, struct ceph_pagelist *pagelist)
{
	struct ceph_osd_data *osd_data;

191
	osd_data = osd_req_op_data(osd_req, which, cls, request_data);
192 193 194 195
	ceph_osd_data_pagelist_init(osd_data, pagelist);
}
EXPORT_SYMBOL(osd_req_op_cls_request_data_pagelist);

196 197 198 199 200 201
void osd_req_op_cls_response_data_pages(struct ceph_osd_request *osd_req,
			unsigned int which, struct page **pages, u64 length,
			u32 alignment, bool pages_from_pool, bool own_pages)
{
	struct ceph_osd_data *osd_data;

202
	osd_data = osd_req_op_data(osd_req, which, cls, response_data);
203 204 205 206 207
	ceph_osd_data_pages_init(osd_data, pages, length, alignment,
				pages_from_pool, own_pages);
}
EXPORT_SYMBOL(osd_req_op_cls_response_data_pages);

208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
static u64 ceph_osd_data_length(struct ceph_osd_data *osd_data)
{
	switch (osd_data->type) {
	case CEPH_OSD_DATA_TYPE_NONE:
		return 0;
	case CEPH_OSD_DATA_TYPE_PAGES:
		return osd_data->length;
	case CEPH_OSD_DATA_TYPE_PAGELIST:
		return (u64)osd_data->pagelist->length;
#ifdef CONFIG_BLOCK
	case CEPH_OSD_DATA_TYPE_BIO:
		return (u64)osd_data->bio_length;
#endif /* CONFIG_BLOCK */
	default:
		WARN(true, "unrecognized data type %d\n", (int)osd_data->type);
		return 0;
	}
}

A
Alex Elder 已提交
227 228
static void ceph_osd_data_release(struct ceph_osd_data *osd_data)
{
229
	if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES && osd_data->own_pages) {
A
Alex Elder 已提交
230 231 232 233 234 235
		int num_pages;

		num_pages = calc_pages_for((u64)osd_data->alignment,
						(u64)osd_data->length);
		ceph_release_page_vector(osd_data->pages, num_pages);
	}
236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253
	ceph_osd_data_init(osd_data);
}

static void osd_req_op_data_release(struct ceph_osd_request *osd_req,
			unsigned int which)
{
	struct ceph_osd_req_op *op;

	BUG_ON(which >= osd_req->r_num_ops);
	op = &osd_req->r_ops[which];

	switch (op->op) {
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
		ceph_osd_data_release(&op->extent.osd_data);
		break;
	case CEPH_OSD_OP_CALL:
		ceph_osd_data_release(&op->cls.request_info);
254
		ceph_osd_data_release(&op->cls.request_data);
255 256 257 258 259
		ceph_osd_data_release(&op->cls.response_data);
		break;
	default:
		break;
	}
A
Alex Elder 已提交
260 261
}

S
Sage Weil 已提交
262 263 264
/*
 * requests
 */
S
Sage Weil 已提交
265
void ceph_osdc_release_request(struct kref *kref)
S
Sage Weil 已提交
266
{
A
Alex Elder 已提交
267
	struct ceph_osd_request *req;
268
	unsigned int which;
S
Sage Weil 已提交
269

A
Alex Elder 已提交
270
	req = container_of(kref, struct ceph_osd_request, r_kref);
S
Sage Weil 已提交
271 272
	if (req->r_request)
		ceph_msg_put(req->r_request);
273
	if (req->r_reply) {
274
		ceph_msg_revoke_incoming(req->r_reply);
275
		ceph_msg_put(req->r_reply);
276
	}
277

278 279
	for (which = 0; which < req->r_num_ops; which++)
		osd_req_op_data_release(req, which);
280

S
Sage Weil 已提交
281 282 283 284 285
	ceph_put_snap_context(req->r_snapc);
	if (req->r_mempool)
		mempool_free(req, req->r_osdc->req_mempool);
	else
		kfree(req);
S
Sage Weil 已提交
286
}
287
EXPORT_SYMBOL(ceph_osdc_release_request);
288

289
struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
S
Sage Weil 已提交
290
					       struct ceph_snap_context *snapc,
291
					       unsigned int num_ops,
292
					       bool use_mempool,
293
					       gfp_t gfp_flags)
S
Sage Weil 已提交
294 295 296
{
	struct ceph_osd_request *req;
	struct ceph_msg *msg;
297 298
	size_t msg_size;

299 300 301
	BUILD_BUG_ON(CEPH_OSD_MAX_OP > U16_MAX);
	BUG_ON(num_ops > CEPH_OSD_MAX_OP);

302 303 304 305 306 307 308 309 310
	msg_size = 4 + 4 + 8 + 8 + 4+8;
	msg_size += 2 + 4 + 8 + 4 + 4; /* oloc */
	msg_size += 1 + 8 + 4 + 4;     /* pg_t */
	msg_size += 4 + MAX_OBJ_NAME_SIZE;
	msg_size += 2 + num_ops*sizeof(struct ceph_osd_op);
	msg_size += 8;  /* snapid */
	msg_size += 8;  /* snap_seq */
	msg_size += 8 * (snapc ? snapc->num_snaps : 0);  /* snaps */
	msg_size += 4;
S
Sage Weil 已提交
311 312

	if (use_mempool) {
313
		req = mempool_alloc(osdc->req_mempool, gfp_flags);
S
Sage Weil 已提交
314 315
		memset(req, 0, sizeof(*req));
	} else {
316
		req = kzalloc(sizeof(*req), gfp_flags);
S
Sage Weil 已提交
317 318
	}
	if (req == NULL)
319
		return NULL;
S
Sage Weil 已提交
320 321 322

	req->r_osdc = osdc;
	req->r_mempool = use_mempool;
323
	req->r_num_ops = num_ops;
324

S
Sage Weil 已提交
325
	kref_init(&req->r_kref);
S
Sage Weil 已提交
326 327
	init_completion(&req->r_completion);
	init_completion(&req->r_safe_completion);
328
	RB_CLEAR_NODE(&req->r_node);
S
Sage Weil 已提交
329
	INIT_LIST_HEAD(&req->r_unsafe_item);
330 331
	INIT_LIST_HEAD(&req->r_linger_item);
	INIT_LIST_HEAD(&req->r_linger_osd);
332
	INIT_LIST_HEAD(&req->r_req_lru_item);
333 334
	INIT_LIST_HEAD(&req->r_osd_item);

S
Sage Weil 已提交
335 336 337 338 339
	/* create reply message */
	if (use_mempool)
		msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
	else
		msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
340
				   OSD_OPREPLY_FRONT_LEN, gfp_flags, true);
341
	if (!msg) {
S
Sage Weil 已提交
342
		ceph_osdc_put_request(req);
343
		return NULL;
S
Sage Weil 已提交
344 345 346 347
	}
	req->r_reply = msg;

	/* create request message; allow space for oid */
S
Sage Weil 已提交
348
	if (use_mempool)
349
		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
S
Sage Weil 已提交
350
	else
351
		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true);
352
	if (!msg) {
S
Sage Weil 已提交
353
		ceph_osdc_put_request(req);
354
		return NULL;
S
Sage Weil 已提交
355
	}
356

S
Sage Weil 已提交
357
	memset(msg->front.iov_base, 0, msg->front.iov_len);
358 359 360 361 362

	req->r_request = msg;

	return req;
}
363
EXPORT_SYMBOL(ceph_osdc_alloc_request);
364

365
static bool osd_req_opcode_valid(u16 opcode)
366
{
367
	switch (opcode) {
368
	case CEPH_OSD_OP_READ:
369
	case CEPH_OSD_OP_STAT:
370 371 372
	case CEPH_OSD_OP_MAPEXT:
	case CEPH_OSD_OP_MASKTRUNC:
	case CEPH_OSD_OP_SPARSE_READ:
373
	case CEPH_OSD_OP_NOTIFY:
374
	case CEPH_OSD_OP_NOTIFY_ACK:
375
	case CEPH_OSD_OP_ASSERT_VER:
376
	case CEPH_OSD_OP_WRITE:
377 378 379 380 381
	case CEPH_OSD_OP_WRITEFULL:
	case CEPH_OSD_OP_TRUNCATE:
	case CEPH_OSD_OP_ZERO:
	case CEPH_OSD_OP_DELETE:
	case CEPH_OSD_OP_APPEND:
382
	case CEPH_OSD_OP_STARTSYNC:
383 384 385 386 387 388
	case CEPH_OSD_OP_SETTRUNC:
	case CEPH_OSD_OP_TRIMTRUNC:
	case CEPH_OSD_OP_TMAPUP:
	case CEPH_OSD_OP_TMAPPUT:
	case CEPH_OSD_OP_TMAPGET:
	case CEPH_OSD_OP_CREATE:
389
	case CEPH_OSD_OP_ROLLBACK:
390
	case CEPH_OSD_OP_WATCH:
391 392 393 394 395 396 397 398 399 400 401 402
	case CEPH_OSD_OP_OMAPGETKEYS:
	case CEPH_OSD_OP_OMAPGETVALS:
	case CEPH_OSD_OP_OMAPGETHEADER:
	case CEPH_OSD_OP_OMAPGETVALSBYKEYS:
	case CEPH_OSD_OP_OMAPSETVALS:
	case CEPH_OSD_OP_OMAPSETHEADER:
	case CEPH_OSD_OP_OMAPCLEAR:
	case CEPH_OSD_OP_OMAPRMKEYS:
	case CEPH_OSD_OP_OMAP_CMP:
	case CEPH_OSD_OP_CLONERANGE:
	case CEPH_OSD_OP_ASSERT_SRC_VERSION:
	case CEPH_OSD_OP_SRC_CMPXATTR:
403
	case CEPH_OSD_OP_GETXATTR:
404
	case CEPH_OSD_OP_GETXATTRS:
405 406
	case CEPH_OSD_OP_CMPXATTR:
	case CEPH_OSD_OP_SETXATTR:
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424
	case CEPH_OSD_OP_SETXATTRS:
	case CEPH_OSD_OP_RESETXATTRS:
	case CEPH_OSD_OP_RMXATTR:
	case CEPH_OSD_OP_PULL:
	case CEPH_OSD_OP_PUSH:
	case CEPH_OSD_OP_BALANCEREADS:
	case CEPH_OSD_OP_UNBALANCEREADS:
	case CEPH_OSD_OP_SCRUB:
	case CEPH_OSD_OP_SCRUB_RESERVE:
	case CEPH_OSD_OP_SCRUB_UNRESERVE:
	case CEPH_OSD_OP_SCRUB_STOP:
	case CEPH_OSD_OP_SCRUB_MAP:
	case CEPH_OSD_OP_WRLOCK:
	case CEPH_OSD_OP_WRUNLOCK:
	case CEPH_OSD_OP_RDLOCK:
	case CEPH_OSD_OP_RDUNLOCK:
	case CEPH_OSD_OP_UPLOCK:
	case CEPH_OSD_OP_DNLOCK:
425
	case CEPH_OSD_OP_CALL:
426 427
	case CEPH_OSD_OP_PGLS:
	case CEPH_OSD_OP_PGLS_FILTER:
428 429 430 431 432 433
		return true;
	default:
		return false;
	}
}

434 435 436 437 438
/*
 * This is an osd op init function for opcodes that have no data or
 * other information associated with them.  It also serves as a
 * common init routine for all the other init functions, below.
 */
439 440 441
static struct ceph_osd_req_op *
osd_req_op_init(struct ceph_osd_request *osd_req, unsigned int which,
				u16 opcode)
442
{
443 444 445
	struct ceph_osd_req_op *op;

	BUG_ON(which >= osd_req->r_num_ops);
446 447
	BUG_ON(!osd_req_opcode_valid(opcode));

448
	op = &osd_req->r_ops[which];
449 450
	memset(op, 0, sizeof (*op));
	op->op = opcode;
451 452

	return op;
453 454
}

455 456
void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
457 458 459
				u64 offset, u64 length,
				u64 truncate_size, u32 truncate_seq)
{
460
	struct ceph_osd_req_op *op = osd_req_op_init(osd_req, which, opcode);
461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
	size_t payload_len = 0;

	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE);

	op->extent.offset = offset;
	op->extent.length = length;
	op->extent.truncate_size = truncate_size;
	op->extent.truncate_seq = truncate_seq;
	if (opcode == CEPH_OSD_OP_WRITE)
		payload_len += length;

	op->payload_len = payload_len;
}
EXPORT_SYMBOL(osd_req_op_extent_init);

476 477
void osd_req_op_extent_update(struct ceph_osd_request *osd_req,
				unsigned int which, u64 length)
478
{
479 480 481 482 483 484
	struct ceph_osd_req_op *op;
	u64 previous;

	BUG_ON(which >= osd_req->r_num_ops);
	op = &osd_req->r_ops[which];
	previous = op->extent.length;
485 486 487 488 489 490 491 492 493 494

	if (length == previous)
		return;		/* Nothing to do */
	BUG_ON(length > previous);

	op->extent.length = length;
	op->payload_len -= previous - length;
}
EXPORT_SYMBOL(osd_req_op_extent_update);

495
void osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which,
496
			u16 opcode, const char *class, const char *method)
497
{
498
	struct ceph_osd_req_op *op = osd_req_op_init(osd_req, which, opcode);
499
	struct ceph_pagelist *pagelist;
500 501 502 503 504
	size_t payload_len = 0;
	size_t size;

	BUG_ON(opcode != CEPH_OSD_OP_CALL);

505 506 507 508
	pagelist = kmalloc(sizeof (*pagelist), GFP_NOFS);
	BUG_ON(!pagelist);
	ceph_pagelist_init(pagelist);

509 510 511 512
	op->cls.class_name = class;
	size = strlen(class);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.class_len = size;
513
	ceph_pagelist_append(pagelist, class, size);
514 515 516 517 518 519
	payload_len += size;

	op->cls.method_name = method;
	size = strlen(method);
	BUG_ON(size > (size_t) U8_MAX);
	op->cls.method_len = size;
520
	ceph_pagelist_append(pagelist, method, size);
521 522
	payload_len += size;

523
	osd_req_op_cls_request_info_pagelist(osd_req, which, pagelist);
524

525 526 527 528 529
	op->cls.argc = 0;	/* currently unused */

	op->payload_len = payload_len;
}
EXPORT_SYMBOL(osd_req_op_cls_init);
530

531 532
void osd_req_op_watch_init(struct ceph_osd_request *osd_req,
				unsigned int which, u16 opcode,
533 534
				u64 cookie, u64 version, int flag)
{
535
	struct ceph_osd_req_op *op = osd_req_op_init(osd_req, which, opcode);
536

537
	BUG_ON(opcode != CEPH_OSD_OP_NOTIFY_ACK && opcode != CEPH_OSD_OP_WATCH);
538 539 540 541 542

	op->watch.cookie = cookie;
	/* op->watch.ver = version; */	/* XXX 3847 */
	op->watch.ver = cpu_to_le64(version);
	if (opcode == CEPH_OSD_OP_WATCH && flag)
543
		op->watch.flag = (u8)1;
544 545 546
}
EXPORT_SYMBOL(osd_req_op_watch_init);

547
static void ceph_osdc_msg_data_add(struct ceph_msg *msg,
548 549 550 551 552 553 554
				struct ceph_osd_data *osd_data)
{
	u64 length = ceph_osd_data_length(osd_data);

	if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
		BUG_ON(length > (u64) SIZE_MAX);
		if (length)
555
			ceph_msg_data_add_pages(msg, osd_data->pages,
556 557 558
					length, osd_data->alignment);
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
		BUG_ON(!length);
559
		ceph_msg_data_add_pagelist(msg, osd_data->pagelist);
560 561
#ifdef CONFIG_BLOCK
	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_BIO) {
562
		ceph_msg_data_add_bio(msg, osd_data->bio, length);
563 564 565 566 567 568
#endif
	} else {
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_NONE);
	}
}

569
static u64 osd_req_encode_op(struct ceph_osd_request *req,
570
			      struct ceph_osd_op *dst, unsigned int which)
571
{
572
	struct ceph_osd_req_op *src;
573
	struct ceph_osd_data *osd_data;
574
	u64 request_data_len = 0;
575
	u64 data_length;
576

577 578
	BUG_ON(which >= req->r_num_ops);
	src = &req->r_ops[which];
579 580 581 582 583 584 585 586 587 588 589 590
	if (WARN_ON(!osd_req_opcode_valid(src->op))) {
		pr_err("unrecognized osd opcode %d\n", src->op);

		return 0;
	}

	switch (src->op) {
	case CEPH_OSD_OP_STAT:
		break;
	case CEPH_OSD_OP_READ:
	case CEPH_OSD_OP_WRITE:
		if (src->op == CEPH_OSD_OP_WRITE)
591
			request_data_len = src->extent.length;
592 593 594 595 596 597
		dst->extent.offset = cpu_to_le64(src->extent.offset);
		dst->extent.length = cpu_to_le64(src->extent.length);
		dst->extent.truncate_size =
			cpu_to_le64(src->extent.truncate_size);
		dst->extent.truncate_seq =
			cpu_to_le32(src->extent.truncate_seq);
598
		osd_data = &src->extent.osd_data;
599
		if (src->op == CEPH_OSD_OP_WRITE)
600
			ceph_osdc_msg_data_add(req->r_request, osd_data);
601
		else
602
			ceph_osdc_msg_data_add(req->r_reply, osd_data);
603 604 605 606
		break;
	case CEPH_OSD_OP_CALL:
		dst->cls.class_len = src->cls.class_len;
		dst->cls.method_len = src->cls.method_len;
607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622
		osd_data = &src->cls.request_info;
		ceph_osdc_msg_data_add(req->r_request, osd_data);
		BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_PAGELIST);
		request_data_len = osd_data->pagelist->length;

		osd_data = &src->cls.request_data;
		data_length = ceph_osd_data_length(osd_data);
		if (data_length) {
			BUG_ON(osd_data->type == CEPH_OSD_DATA_TYPE_NONE);
			dst->cls.indata_len = cpu_to_le32(data_length);
			ceph_osdc_msg_data_add(req->r_request, osd_data);
			src->payload_len += data_length;
			request_data_len += data_length;
		}
		osd_data = &src->cls.response_data;
		ceph_osdc_msg_data_add(req->r_reply, osd_data);
623 624 625 626 627 628 629 630 631 632
		break;
	case CEPH_OSD_OP_STARTSYNC:
		break;
	case CEPH_OSD_OP_NOTIFY_ACK:
	case CEPH_OSD_OP_WATCH:
		dst->watch.cookie = cpu_to_le64(src->watch.cookie);
		dst->watch.ver = cpu_to_le64(src->watch.ver);
		dst->watch.flag = src->watch.flag;
		break;
	default:
633
		pr_err("unsupported osd opcode %s\n",
634
			ceph_osd_op_name(src->op));
635
		WARN_ON(1);
636 637

		return 0;
638
	}
639
	dst->op = cpu_to_le16(src->op);
640
	dst->payload_len = cpu_to_le32(src->payload_len);
641

642
	return request_data_len;
643 644
}

645 646 647 648 649 650 651 652 653 654 655 656 657 658
/*
 * build new request AND message, calculate layout, and adjust file
 * extent as needed.
 *
 * if the file was recently truncated, we include information about its
 * old and new size so that the object can be updated appropriately.  (we
 * avoid synchronously deleting truncated objects because it's slow.)
 *
 * if @do_sync, include a 'startsync' command so that the osd will flush
 * data quickly.
 */
struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
					       struct ceph_file_layout *layout,
					       struct ceph_vino vino,
659
					       u64 off, u64 *plen, int num_ops,
660 661 662 663
					       int opcode, int flags,
					       struct ceph_snap_context *snapc,
					       u32 truncate_seq,
					       u64 truncate_size,
664
					       bool use_mempool)
665
{
666
	struct ceph_osd_request *req;
667 668 669
	u64 objnum = 0;
	u64 objoff = 0;
	u64 objlen = 0;
670 671
	u32 object_size;
	u64 object_base;
672
	int r;
673

674
	BUG_ON(opcode != CEPH_OSD_OP_READ && opcode != CEPH_OSD_OP_WRITE);
675

676
	req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool,
A
Alex Elder 已提交
677
					GFP_NOFS);
678
	if (!req)
679
		return ERR_PTR(-ENOMEM);
680

681
	req->r_flags = flags;
682 683

	/* calculate max write size */
684
	r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen);
685 686
	if (r < 0) {
		ceph_osdc_put_request(req);
687
		return ERR_PTR(r);
688
	}
689

690 691 692 693 694 695 696 697
	object_size = le32_to_cpu(layout->fl_object_size);
	object_base = off - objoff;
	if (truncate_size <= object_base) {
		truncate_size = 0;
	} else {
		truncate_size -= object_base;
		if (truncate_size > object_size)
			truncate_size = object_size;
698
	}
699

700
	osd_req_op_extent_init(req, 0, opcode, objoff, objlen,
701
				truncate_size, truncate_seq);
702

703 704 705 706 707 708
	/*
	 * A second op in the ops array means the caller wants to
	 * also issue a include a 'startsync' command so that the
	 * osd will flush data quickly.
	 */
	if (num_ops > 1)
709
		osd_req_op_init(req, 1, CEPH_OSD_OP_STARTSYNC);
710

711 712
	req->r_file_layout = *layout;  /* keep a copy */

713 714
	snprintf(req->r_oid, sizeof(req->r_oid), "%llx.%08llx",
		vino.ino, objnum);
715 716
	req->r_oid_len = strlen(req->r_oid);

S
Sage Weil 已提交
717 718
	return req;
}
719
EXPORT_SYMBOL(ceph_osdc_new_request);
S
Sage Weil 已提交
720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785

/*
 * We keep osd requests in an rbtree, sorted by ->r_tid.
 */
static void __insert_request(struct ceph_osd_client *osdc,
			     struct ceph_osd_request *new)
{
	struct rb_node **p = &osdc->requests.rb_node;
	struct rb_node *parent = NULL;
	struct ceph_osd_request *req = NULL;

	while (*p) {
		parent = *p;
		req = rb_entry(parent, struct ceph_osd_request, r_node);
		if (new->r_tid < req->r_tid)
			p = &(*p)->rb_left;
		else if (new->r_tid > req->r_tid)
			p = &(*p)->rb_right;
		else
			BUG();
	}

	rb_link_node(&new->r_node, parent, p);
	rb_insert_color(&new->r_node, &osdc->requests);
}

static struct ceph_osd_request *__lookup_request(struct ceph_osd_client *osdc,
						 u64 tid)
{
	struct ceph_osd_request *req;
	struct rb_node *n = osdc->requests.rb_node;

	while (n) {
		req = rb_entry(n, struct ceph_osd_request, r_node);
		if (tid < req->r_tid)
			n = n->rb_left;
		else if (tid > req->r_tid)
			n = n->rb_right;
		else
			return req;
	}
	return NULL;
}

static struct ceph_osd_request *
__lookup_request_ge(struct ceph_osd_client *osdc,
		    u64 tid)
{
	struct ceph_osd_request *req;
	struct rb_node *n = osdc->requests.rb_node;

	while (n) {
		req = rb_entry(n, struct ceph_osd_request, r_node);
		if (tid < req->r_tid) {
			if (!n->rb_left)
				return req;
			n = n->rb_left;
		} else if (tid > req->r_tid) {
			n = n->rb_right;
		} else {
			return req;
		}
	}
	return NULL;
}

786 787 788 789 790 791
/*
 * Resubmit requests pending on the given osd.
 */
static void __kick_osd_requests(struct ceph_osd_client *osdc,
				struct ceph_osd *osd)
{
792
	struct ceph_osd_request *req, *nreq;
793
	LIST_HEAD(resend);
794 795 796 797
	int err;

	dout("__kick_osd_requests osd%d\n", osd->o_osd);
	err = __reset_osd(osdc, osd);
798
	if (err)
799
		return;
800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816
	/*
	 * Build up a list of requests to resend by traversing the
	 * osd's list of requests.  Requests for a given object are
	 * sent in tid order, and that is also the order they're
	 * kept on this list.  Therefore all requests that are in
	 * flight will be found first, followed by all requests that
	 * have not yet been sent.  And to resend requests while
	 * preserving this order we will want to put any sent
	 * requests back on the front of the osd client's unsent
	 * list.
	 *
	 * So we build a separate ordered list of already-sent
	 * requests for the affected osd and splice it onto the
	 * front of the osd client's unsent list.  Once we've seen a
	 * request that has not yet been sent we're done.  Those
	 * requests are already sitting right where they belong.
	 */
817
	list_for_each_entry(req, &osd->o_requests, r_osd_item) {
818 819 820 821
		if (!req->r_sent)
			break;
		list_move_tail(&req->r_req_lru_item, &resend);
		dout("requeueing %p tid %llu osd%d\n", req, req->r_tid,
822
		     osd->o_osd);
823 824 825
		if (!req->r_linger)
			req->r_flags |= CEPH_OSD_FLAG_RETRY;
	}
826
	list_splice(&resend, &osdc->req_unsent);
827

828 829 830 831 832
	/*
	 * Linger requests are re-registered before sending, which
	 * sets up a new tid for each.  We add them to the unsent
	 * list at the end to keep things in tid order.
	 */
833 834
	list_for_each_entry_safe(req, nreq, &osd->o_linger_requests,
				 r_linger_osd) {
835 836 837 838 839
		/*
		 * reregister request prior to unregistering linger so
		 * that r_osd is preserved.
		 */
		BUG_ON(!list_empty(&req->r_req_lru_item));
840
		__register_request(osdc, req);
841
		list_add_tail(&req->r_req_lru_item, &osdc->req_unsent);
842
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
843
		__unregister_linger_request(osdc, req);
844 845
		dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid,
		     osd->o_osd);
846 847 848
	}
}

S
Sage Weil 已提交
849
/*
850
 * If the osd connection drops, we need to resubmit all requests.
S
Sage Weil 已提交
851 852 853 854 855 856 857 858 859 860 861
 */
static void osd_reset(struct ceph_connection *con)
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc;

	if (!osd)
		return;
	dout("osd_reset osd%d\n", osd->o_osd);
	osdc = osd->o_osdc;
	down_read(&osdc->map_sem);
S
Sage Weil 已提交
862 863
	mutex_lock(&osdc->request_mutex);
	__kick_osd_requests(osdc, osd);
A
Alex Elder 已提交
864
	__send_queued(osdc);
S
Sage Weil 已提交
865
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
866 867 868 869 870 871
	up_read(&osdc->map_sem);
}

/*
 * Track open sessions with osds.
 */
872
static struct ceph_osd *create_osd(struct ceph_osd_client *osdc, int onum)
S
Sage Weil 已提交
873 874 875 876 877 878 879 880 881
{
	struct ceph_osd *osd;

	osd = kzalloc(sizeof(*osd), GFP_NOFS);
	if (!osd)
		return NULL;

	atomic_set(&osd->o_ref, 1);
	osd->o_osdc = osdc;
882
	osd->o_osd = onum;
883
	RB_CLEAR_NODE(&osd->o_node);
S
Sage Weil 已提交
884
	INIT_LIST_HEAD(&osd->o_requests);
885
	INIT_LIST_HEAD(&osd->o_linger_requests);
886
	INIT_LIST_HEAD(&osd->o_osd_lru);
S
Sage Weil 已提交
887 888
	osd->o_incarnation = 1;

889
	ceph_con_init(&osd->o_con, osd, &osd_con_ops, &osdc->client->msgr);
890

891
	INIT_LIST_HEAD(&osd->o_keepalive_item);
S
Sage Weil 已提交
892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910
	return osd;
}

static struct ceph_osd *get_osd(struct ceph_osd *osd)
{
	if (atomic_inc_not_zero(&osd->o_ref)) {
		dout("get_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref)-1,
		     atomic_read(&osd->o_ref));
		return osd;
	} else {
		dout("get_osd %p FAIL\n", osd);
		return NULL;
	}
}

static void put_osd(struct ceph_osd *osd)
{
	dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref),
	     atomic_read(&osd->o_ref) - 1);
911
	if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) {
S
Sage Weil 已提交
912 913
		struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;

914
		ceph_auth_destroy_authorizer(ac, osd->o_auth.authorizer);
S
Sage Weil 已提交
915
		kfree(osd);
S
Sage Weil 已提交
916
	}
S
Sage Weil 已提交
917 918 919 920 921
}

/*
 * remove an osd from our map
 */
922
static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
923
{
924
	dout("__remove_osd %p\n", osd);
S
Sage Weil 已提交
925 926
	BUG_ON(!list_empty(&osd->o_requests));
	rb_erase(&osd->o_node, &osdc->osds);
927
	list_del_init(&osd->o_osd_lru);
S
Sage Weil 已提交
928 929 930 931
	ceph_con_close(&osd->o_con);
	put_osd(osd);
}

932 933
static void remove_all_osds(struct ceph_osd_client *osdc)
{
934
	dout("%s %p\n", __func__, osdc);
935 936 937 938 939 940 941 942 943
	mutex_lock(&osdc->request_mutex);
	while (!RB_EMPTY_ROOT(&osdc->osds)) {
		struct ceph_osd *osd = rb_entry(rb_first(&osdc->osds),
						struct ceph_osd, o_node);
		__remove_osd(osdc, osd);
	}
	mutex_unlock(&osdc->request_mutex);
}

944 945 946 947 948 949
static void __move_osd_to_lru(struct ceph_osd_client *osdc,
			      struct ceph_osd *osd)
{
	dout("__move_osd_to_lru %p\n", osd);
	BUG_ON(!list_empty(&osd->o_osd_lru));
	list_add_tail(&osd->o_osd_lru, &osdc->osd_lru);
950
	osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ;
951 952 953 954 955 956 957 958 959
}

static void __remove_osd_from_lru(struct ceph_osd *osd)
{
	dout("__remove_osd_from_lru %p\n", osd);
	if (!list_empty(&osd->o_osd_lru))
		list_del_init(&osd->o_osd_lru);
}

960
static void remove_old_osds(struct ceph_osd_client *osdc)
961 962 963 964 965 966
{
	struct ceph_osd *osd, *nosd;

	dout("__remove_old_osds %p\n", osdc);
	mutex_lock(&osdc->request_mutex);
	list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) {
967
		if (time_before(jiffies, osd->lru_ttl))
968 969 970 971 972 973
			break;
		__remove_osd(osdc, osd);
	}
	mutex_unlock(&osdc->request_mutex);
}

S
Sage Weil 已提交
974 975 976
/*
 * reset osd connect
 */
977
static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd)
S
Sage Weil 已提交
978
{
A
Alex Elder 已提交
979
	struct ceph_entity_addr *peer_addr;
S
Sage Weil 已提交
980

981
	dout("__reset_osd %p osd%d\n", osd, osd->o_osd);
982 983
	if (list_empty(&osd->o_requests) &&
	    list_empty(&osd->o_linger_requests)) {
984
		__remove_osd(osdc, osd);
A
Alex Elder 已提交
985 986 987 988 989 990 991 992 993

		return -ENODEV;
	}

	peer_addr = &osdc->osdmap->osd_addr[osd->o_osd];
	if (!memcmp(peer_addr, &osd->o_con.peer_addr, sizeof (*peer_addr)) &&
			!ceph_con_opened(&osd->o_con)) {
		struct ceph_osd_request *req;

994 995 996 997 998
		dout(" osd addr hasn't changed and connection never opened,"
		     " letting msgr retry");
		/* touch each r_stamp for handle_timeout()'s benfit */
		list_for_each_entry(req, &osd->o_requests, r_osd_item)
			req->r_stamp = jiffies;
A
Alex Elder 已提交
999 1000

		return -EAGAIN;
S
Sage Weil 已提交
1001
	}
A
Alex Elder 已提交
1002 1003 1004 1005 1006 1007

	ceph_con_close(&osd->o_con);
	ceph_con_open(&osd->o_con, CEPH_ENTITY_TYPE_OSD, osd->o_osd, peer_addr);
	osd->o_incarnation++;

	return 0;
S
Sage Weil 已提交
1008 1009 1010 1011 1012 1013 1014 1015
}

static void __insert_osd(struct ceph_osd_client *osdc, struct ceph_osd *new)
{
	struct rb_node **p = &osdc->osds.rb_node;
	struct rb_node *parent = NULL;
	struct ceph_osd *osd = NULL;

1016
	dout("__insert_osd %p osd%d\n", new, new->o_osd);
S
Sage Weil 已提交
1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048
	while (*p) {
		parent = *p;
		osd = rb_entry(parent, struct ceph_osd, o_node);
		if (new->o_osd < osd->o_osd)
			p = &(*p)->rb_left;
		else if (new->o_osd > osd->o_osd)
			p = &(*p)->rb_right;
		else
			BUG();
	}

	rb_link_node(&new->o_node, parent, p);
	rb_insert_color(&new->o_node, &osdc->osds);
}

static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o)
{
	struct ceph_osd *osd;
	struct rb_node *n = osdc->osds.rb_node;

	while (n) {
		osd = rb_entry(n, struct ceph_osd, o_node);
		if (o < osd->o_osd)
			n = n->rb_left;
		else if (o > osd->o_osd)
			n = n->rb_right;
		else
			return osd;
	}
	return NULL;
}

1049 1050 1051
static void __schedule_osd_timeout(struct ceph_osd_client *osdc)
{
	schedule_delayed_work(&osdc->timeout_work,
1052
			osdc->client->options->osd_keepalive_timeout * HZ);
1053 1054 1055 1056 1057 1058
}

static void __cancel_osd_timeout(struct ceph_osd_client *osdc)
{
	cancel_delayed_work(&osdc->timeout_work);
}
S
Sage Weil 已提交
1059 1060 1061 1062 1063

/*
 * Register request, assign tid.  If this is the first request, set up
 * the timeout event.
 */
1064 1065
static void __register_request(struct ceph_osd_client *osdc,
			       struct ceph_osd_request *req)
S
Sage Weil 已提交
1066 1067
{
	req->r_tid = ++osdc->last_tid;
1068
	req->r_request->hdr.tid = cpu_to_le64(req->r_tid);
1069
	dout("__register_request %p tid %lld\n", req, req->r_tid);
S
Sage Weil 已提交
1070 1071 1072 1073
	__insert_request(osdc, req);
	ceph_osdc_get_request(req);
	osdc->num_requests++;
	if (osdc->num_requests == 1) {
1074 1075
		dout(" first request, scheduling timeout\n");
		__schedule_osd_timeout(osdc);
S
Sage Weil 已提交
1076
	}
1077 1078
}

S
Sage Weil 已提交
1079 1080 1081 1082 1083 1084
/*
 * called under osdc->request_mutex
 */
static void __unregister_request(struct ceph_osd_client *osdc,
				 struct ceph_osd_request *req)
{
1085 1086 1087 1088 1089 1090
	if (RB_EMPTY_NODE(&req->r_node)) {
		dout("__unregister_request %p tid %lld not registered\n",
			req, req->r_tid);
		return;
	}

S
Sage Weil 已提交
1091 1092 1093 1094
	dout("__unregister_request %p tid %lld\n", req, req->r_tid);
	rb_erase(&req->r_node, &osdc->requests);
	osdc->num_requests--;

1095 1096
	if (req->r_osd) {
		/* make sure the original request isn't in flight. */
1097
		ceph_msg_revoke(req->r_request);
1098 1099

		list_del_init(&req->r_osd_item);
1100 1101 1102
		if (list_empty(&req->r_osd->o_requests) &&
		    list_empty(&req->r_osd->o_linger_requests)) {
			dout("moving osd to %p lru\n", req->r_osd);
1103
			__move_osd_to_lru(osdc, req->r_osd);
1104
		}
1105
		if (list_empty(&req->r_linger_item))
1106
			req->r_osd = NULL;
1107
	}
S
Sage Weil 已提交
1108

A
Alex Elder 已提交
1109
	list_del_init(&req->r_req_lru_item);
S
Sage Weil 已提交
1110 1111
	ceph_osdc_put_request(req);

1112 1113 1114
	if (osdc->num_requests == 0) {
		dout(" no requests, canceling timeout\n");
		__cancel_osd_timeout(osdc);
S
Sage Weil 已提交
1115 1116 1117 1118 1119 1120 1121 1122
	}
}

/*
 * Cancel a previously queued request message
 */
static void __cancel_request(struct ceph_osd_request *req)
{
1123
	if (req->r_sent && req->r_osd) {
1124
		ceph_msg_revoke(req->r_request);
S
Sage Weil 已提交
1125 1126 1127 1128
		req->r_sent = 0;
	}
}

1129 1130 1131 1132 1133
static void __register_linger_request(struct ceph_osd_client *osdc,
				    struct ceph_osd_request *req)
{
	dout("__register_linger_request %p\n", req);
	list_add_tail(&req->r_linger_item, &osdc->req_linger);
1134 1135 1136
	if (req->r_osd)
		list_add_tail(&req->r_linger_osd,
			      &req->r_osd->o_linger_requests);
1137 1138 1139 1140 1141 1142
}

static void __unregister_linger_request(struct ceph_osd_client *osdc,
					struct ceph_osd_request *req)
{
	dout("__unregister_linger_request %p\n", req);
A
Alex Elder 已提交
1143
	list_del_init(&req->r_linger_item);
1144 1145 1146 1147 1148 1149 1150 1151
	if (req->r_osd) {
		list_del_init(&req->r_linger_osd);

		if (list_empty(&req->r_osd->o_requests) &&
		    list_empty(&req->r_osd->o_linger_requests)) {
			dout("moving osd to %p lru\n", req->r_osd);
			__move_osd_to_lru(osdc, req->r_osd);
		}
1152 1153
		if (list_empty(&req->r_osd_item))
			req->r_osd = NULL;
1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183
	}
}

void ceph_osdc_unregister_linger_request(struct ceph_osd_client *osdc,
					 struct ceph_osd_request *req)
{
	mutex_lock(&osdc->request_mutex);
	if (req->r_linger) {
		__unregister_linger_request(osdc, req);
		ceph_osdc_put_request(req);
	}
	mutex_unlock(&osdc->request_mutex);
}
EXPORT_SYMBOL(ceph_osdc_unregister_linger_request);

void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc,
				  struct ceph_osd_request *req)
{
	if (!req->r_linger) {
		dout("set_request_linger %p\n", req);
		req->r_linger = 1;
		/*
		 * caller is now responsible for calling
		 * unregister_linger_request
		 */
		ceph_osdc_get_request(req);
	}
}
EXPORT_SYMBOL(ceph_osdc_set_request_linger);

S
Sage Weil 已提交
1184 1185 1186
/*
 * Pick an osd (the first 'up' osd in the pg), allocate the osd struct
 * (as needed), and set the request r_osd appropriately.  If there is
L
Lucas De Marchi 已提交
1187
 * no up osd, set r_osd to NULL.  Move the request to the appropriate list
1188
 * (unsent, homeless) or leave on in-flight lru.
S
Sage Weil 已提交
1189 1190 1191 1192 1193
 *
 * Return 0 if unchanged, 1 if changed, or negative on error.
 *
 * Caller should hold map_sem for read and request_mutex.
 */
1194
static int __map_request(struct ceph_osd_client *osdc,
1195
			 struct ceph_osd_request *req, int force_resend)
S
Sage Weil 已提交
1196
{
1197
	struct ceph_pg pgid;
1198 1199
	int acting[CEPH_PG_MAX_SIZE];
	int o = -1, num = 0;
S
Sage Weil 已提交
1200 1201
	int err;

1202
	dout("map_request %p tid %lld\n", req, req->r_tid);
1203 1204
	err = ceph_calc_ceph_pg(&pgid, req->r_oid, osdc->osdmap,
				ceph_file_layout_pg_pool(req->r_file_layout));
1205 1206
	if (err) {
		list_move(&req->r_req_lru_item, &osdc->req_notarget);
S
Sage Weil 已提交
1207
		return err;
1208
	}
1209 1210
	req->r_pgid = pgid;

1211 1212 1213 1214 1215
	err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
	if (err > 0) {
		o = acting[0];
		num = err;
	}
S
Sage Weil 已提交
1216

1217 1218
	if ((!force_resend &&
	     req->r_osd && req->r_osd->o_osd == o &&
1219 1220 1221
	     req->r_sent >= req->r_osd->o_incarnation &&
	     req->r_num_pg_osds == num &&
	     memcmp(req->r_pg_osds, acting, sizeof(acting[0])*num) == 0) ||
S
Sage Weil 已提交
1222 1223 1224
	    (req->r_osd == NULL && o == -1))
		return 0;  /* no change */

1225 1226
	dout("map_request tid %llu pgid %lld.%x osd%d (was osd%d)\n",
	     req->r_tid, pgid.pool, pgid.seed, o,
S
Sage Weil 已提交
1227 1228
	     req->r_osd ? req->r_osd->o_osd : -1);

1229 1230 1231 1232
	/* record full pg acting set */
	memcpy(req->r_pg_osds, acting, sizeof(acting[0]) * num);
	req->r_num_pg_osds = num;

S
Sage Weil 已提交
1233 1234 1235 1236 1237 1238 1239 1240
	if (req->r_osd) {
		__cancel_request(req);
		list_del_init(&req->r_osd_item);
		req->r_osd = NULL;
	}

	req->r_osd = __lookup_osd(osdc, o);
	if (!req->r_osd && o >= 0) {
1241
		err = -ENOMEM;
1242
		req->r_osd = create_osd(osdc, o);
1243 1244
		if (!req->r_osd) {
			list_move(&req->r_req_lru_item, &osdc->req_notarget);
1245
			goto out;
1246
		}
S
Sage Weil 已提交
1247

1248
		dout("map_request osd %p is osd%d\n", req->r_osd, o);
S
Sage Weil 已提交
1249 1250
		__insert_osd(osdc, req->r_osd);

1251 1252 1253
		ceph_con_open(&req->r_osd->o_con,
			      CEPH_ENTITY_TYPE_OSD, o,
			      &osdc->osdmap->osd_addr[o]);
S
Sage Weil 已提交
1254 1255
	}

1256 1257
	if (req->r_osd) {
		__remove_osd_from_lru(req->r_osd);
1258 1259
		list_add_tail(&req->r_osd_item, &req->r_osd->o_requests);
		list_move_tail(&req->r_req_lru_item, &osdc->req_unsent);
1260
	} else {
1261
		list_move_tail(&req->r_req_lru_item, &osdc->req_notarget);
1262
	}
1263
	err = 1;   /* osd or pg changed */
S
Sage Weil 已提交
1264 1265 1266 1267 1268 1269 1270 1271

out:
	return err;
}

/*
 * caller should hold map_sem (for read) and request_mutex
 */
1272 1273
static void __send_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
S
Sage Weil 已提交
1274
{
1275
	void *p;
S
Sage Weil 已提交
1276

1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
	dout("send_request %p tid %llu to osd%d flags %d pg %lld.%x\n",
	     req, req->r_tid, req->r_osd->o_osd, req->r_flags,
	     (unsigned long long)req->r_pgid.pool, req->r_pgid.seed);

	/* fill in message content that changes each time we send it */
	put_unaligned_le32(osdc->osdmap->epoch, req->r_request_osdmap_epoch);
	put_unaligned_le32(req->r_flags, req->r_request_flags);
	put_unaligned_le64(req->r_pgid.pool, req->r_request_pool);
	p = req->r_request_pgid;
	ceph_encode_64(&p, req->r_pgid.pool);
	ceph_encode_32(&p, req->r_pgid.seed);
	put_unaligned_le64(1, req->r_request_attempts);  /* FIXME */
	memcpy(req->r_request_reassert_version, &req->r_reassert_version,
	       sizeof(req->r_reassert_version));
1291

S
Sage Weil 已提交
1292
	req->r_stamp = jiffies;
1293
	list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
S
Sage Weil 已提交
1294 1295

	ceph_msg_get(req->r_request); /* send consumes a ref */
1296 1297 1298 1299 1300

	/* Mark the request unsafe if this is the first timet's being sent. */

	if (!req->r_sent && req->r_unsafe_callback)
		req->r_unsafe_callback(req, true);
S
Sage Weil 已提交
1301
	req->r_sent = req->r_osd->o_incarnation;
1302 1303

	ceph_con_send(&req->r_osd->o_con, req->r_request);
S
Sage Weil 已提交
1304 1305
}

1306 1307 1308
/*
 * Send any requests in the queue (req_unsent).
 */
A
Alex Elder 已提交
1309
static void __send_queued(struct ceph_osd_client *osdc)
1310 1311 1312
{
	struct ceph_osd_request *req, *tmp;

A
Alex Elder 已提交
1313 1314
	dout("__send_queued\n");
	list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item)
1315 1316 1317
		__send_request(osdc, req);
}

S
Sage Weil 已提交
1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330
/*
 * Timeout callback, called every N seconds when 1 or more osd
 * requests has been active for more than N seconds.  When this
 * happens, we ping all OSDs with requests who have timed out to
 * ensure any communications channel reset is detected.  Reset the
 * request timeouts another N seconds in the future as we go.
 * Reschedule the timeout event another N seconds in future (unless
 * there are no open requests).
 */
static void handle_timeout(struct work_struct *work)
{
	struct ceph_osd_client *osdc =
		container_of(work, struct ceph_osd_client, timeout_work.work);
S
Sage Weil 已提交
1331
	struct ceph_osd_request *req;
S
Sage Weil 已提交
1332
	struct ceph_osd *osd;
1333
	unsigned long keepalive =
1334
		osdc->client->options->osd_keepalive_timeout * HZ;
1335
	struct list_head slow_osds;
S
Sage Weil 已提交
1336 1337 1338 1339 1340 1341 1342
	dout("timeout\n");
	down_read(&osdc->map_sem);

	ceph_monc_request_next_osdmap(&osdc->client->monc);

	mutex_lock(&osdc->request_mutex);

1343 1344 1345 1346 1347 1348 1349
	/*
	 * ping osds that are a bit slow.  this ensures that if there
	 * is a break in the TCP connection we will notice, and reopen
	 * a connection with that osd (from the fault callback).
	 */
	INIT_LIST_HEAD(&slow_osds);
	list_for_each_entry(req, &osdc->req_lru, r_req_lru_item) {
S
Sage Weil 已提交
1350
		if (time_before(jiffies, req->r_stamp + keepalive))
1351 1352 1353 1354 1355
			break;

		osd = req->r_osd;
		BUG_ON(!osd);
		dout(" tid %llu is slow, will send keepalive on osd%d\n",
S
Sage Weil 已提交
1356
		     req->r_tid, osd->o_osd);
1357 1358 1359 1360 1361 1362
		list_move_tail(&osd->o_keepalive_item, &slow_osds);
	}
	while (!list_empty(&slow_osds)) {
		osd = list_entry(slow_osds.next, struct ceph_osd,
				 o_keepalive_item);
		list_del_init(&osd->o_keepalive_item);
S
Sage Weil 已提交
1363 1364 1365
		ceph_con_keepalive(&osd->o_con);
	}

1366
	__schedule_osd_timeout(osdc);
A
Alex Elder 已提交
1367
	__send_queued(osdc);
S
Sage Weil 已提交
1368 1369 1370 1371
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
}

1372 1373 1374 1375 1376 1377
static void handle_osds_timeout(struct work_struct *work)
{
	struct ceph_osd_client *osdc =
		container_of(work, struct ceph_osd_client,
			     osds_timeout_work.work);
	unsigned long delay =
1378
		osdc->client->options->osd_idle_ttl * HZ >> 2;
1379 1380 1381

	dout("osds timeout\n");
	down_read(&osdc->map_sem);
1382
	remove_old_osds(osdc);
1383 1384 1385 1386 1387 1388
	up_read(&osdc->map_sem);

	schedule_delayed_work(&osdc->osds_timeout_work,
			      round_jiffies_relative(delay));
}

S
Sage Weil 已提交
1389 1390
static void complete_request(struct ceph_osd_request *req)
{
1391 1392
	if (req->r_unsafe_callback)
		req->r_unsafe_callback(req, false);
S
Sage Weil 已提交
1393 1394 1395
	complete_all(&req->r_safe_completion);  /* fsync waiter */
}

S
Sage Weil 已提交
1396 1397 1398 1399
/*
 * handle osd op reply.  either call the callback if it is specified,
 * or do the completion to wake up the waiting thread.
 */
1400 1401
static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
			 struct ceph_connection *con)
S
Sage Weil 已提交
1402
{
1403
	void *p, *end;
S
Sage Weil 已提交
1404 1405
	struct ceph_osd_request *req;
	u64 tid;
1406
	int object_len;
1407 1408
	unsigned int numops;
	int payload_len, flags;
1409
	s32 result;
1410 1411 1412 1413 1414 1415
	s32 retry_attempt;
	struct ceph_pg pg;
	int err;
	u32 reassert_epoch;
	u64 reassert_version;
	u32 osdmap_epoch;
1416
	int already_completed;
1417
	u32 bytes;
1418
	unsigned int i;
S
Sage Weil 已提交
1419

1420
	tid = le64_to_cpu(msg->hdr.tid);
1421 1422 1423 1424 1425 1426 1427 1428 1429 1430
	dout("handle_reply %p tid %llu\n", msg, tid);

	p = msg->front.iov_base;
	end = p + msg->front.iov_len;

	ceph_decode_need(&p, end, 4, bad);
	object_len = ceph_decode_32(&p);
	ceph_decode_need(&p, end, object_len, bad);
	p += object_len;

1431
	err = ceph_decode_pgid(&p, end, &pg);
1432
	if (err)
S
Sage Weil 已提交
1433
		goto bad;
1434 1435 1436 1437 1438 1439 1440 1441

	ceph_decode_need(&p, end, 8 + 4 + 4 + 8 + 4, bad);
	flags = ceph_decode_64(&p);
	result = ceph_decode_32(&p);
	reassert_epoch = ceph_decode_32(&p);
	reassert_version = ceph_decode_64(&p);
	osdmap_epoch = ceph_decode_32(&p);

S
Sage Weil 已提交
1442 1443 1444 1445 1446
	/* lookup */
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (req == NULL) {
		dout("handle_reply tid %llu dne\n", tid);
1447
		goto bad_mutex;
S
Sage Weil 已提交
1448 1449
	}
	ceph_osdc_get_request(req);
1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471

	dout("handle_reply %p tid %llu req %p result %d\n", msg, tid,
	     req, result);

	ceph_decode_need(&p, end, 4, bad);
	numops = ceph_decode_32(&p);
	if (numops > CEPH_OSD_MAX_OP)
		goto bad_put;
	if (numops != req->r_num_ops)
		goto bad_put;
	payload_len = 0;
	ceph_decode_need(&p, end, numops * sizeof(struct ceph_osd_op), bad);
	for (i = 0; i < numops; i++) {
		struct ceph_osd_op *op = p;
		int len;

		len = le32_to_cpu(op->payload_len);
		req->r_reply_op_len[i] = len;
		dout(" op %d has %d bytes\n", i, len);
		payload_len += len;
		p += sizeof(*op);
	}
1472 1473
	bytes = le32_to_cpu(msg->hdr.data_len);
	if (payload_len != bytes) {
1474
		pr_warning("sum of op payload lens %d != data_len %d",
1475
			   payload_len, bytes);
1476 1477 1478 1479 1480 1481 1482
		goto bad_put;
	}

	ceph_decode_need(&p, end, 4 + numops * 4, bad);
	retry_attempt = ceph_decode_32(&p);
	for (i = 0; i < numops; i++)
		req->r_reply_op_result[i] = ceph_decode_32(&p);
S
Sage Weil 已提交
1483 1484 1485

	if (!req->r_got_reply) {

1486
		req->r_result = result;
S
Sage Weil 已提交
1487 1488 1489 1490 1491 1492
		dout("handle_reply result %d bytes %d\n", req->r_result,
		     bytes);
		if (req->r_result == 0)
			req->r_result = bytes;

		/* in case this is a write and we need to replay, */
1493 1494
		req->r_reassert_version.epoch = cpu_to_le32(reassert_epoch);
		req->r_reassert_version.version = cpu_to_le64(reassert_version);
S
Sage Weil 已提交
1495 1496 1497 1498

		req->r_got_reply = 1;
	} else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) {
		dout("handle_reply tid %llu dup ack\n", tid);
S
Sage Weil 已提交
1499
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1500 1501 1502 1503 1504
		goto done;
	}

	dout("handle_reply tid %llu flags %d\n", tid, flags);

1505 1506 1507
	if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK))
		__register_linger_request(osdc, req);

S
Sage Weil 已提交
1508
	/* either this is a read, or we got the safe response */
1509 1510
	if (result < 0 ||
	    (flags & CEPH_OSD_FLAG_ONDISK) ||
S
Sage Weil 已提交
1511 1512 1513
	    ((flags & CEPH_OSD_FLAG_WRITE) == 0))
		__unregister_request(osdc, req);

1514 1515
	already_completed = req->r_completed;
	req->r_completed = 1;
S
Sage Weil 已提交
1516
	mutex_unlock(&osdc->request_mutex);
1517 1518
	if (already_completed)
		goto done;
S
Sage Weil 已提交
1519 1520 1521 1522

	if (req->r_callback)
		req->r_callback(req, msg);
	else
1523
		complete_all(&req->r_completion);
S
Sage Weil 已提交
1524

S
Sage Weil 已提交
1525 1526
	if (flags & CEPH_OSD_FLAG_ONDISK)
		complete_request(req);
S
Sage Weil 已提交
1527 1528

done:
1529
	dout("req=%p req->r_linger=%d\n", req, req->r_linger);
S
Sage Weil 已提交
1530 1531 1532
	ceph_osdc_put_request(req);
	return;

1533 1534
bad_put:
	ceph_osdc_put_request(req);
1535 1536
bad_mutex:
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1537
bad:
1538 1539
	pr_err("corrupt osd_op_reply got %d %d\n",
	       (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len));
1540
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1541 1542
}

1543
static void reset_changed_osds(struct ceph_osd_client *osdc)
S
Sage Weil 已提交
1544 1545 1546
{
	struct rb_node *p, *n;

1547 1548
	for (p = rb_first(&osdc->osds); p; p = n) {
		struct ceph_osd *osd = rb_entry(p, struct ceph_osd, o_node);
S
Sage Weil 已提交
1549

1550 1551 1552 1553 1554 1555 1556
		n = rb_next(p);
		if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) ||
		    memcmp(&osd->o_con.peer_addr,
			   ceph_osd_addr(osdc->osdmap,
					 osd->o_osd),
			   sizeof(struct ceph_entity_addr)) != 0)
			__reset_osd(osdc, osd);
S
Sage Weil 已提交
1557
	}
1558 1559 1560
}

/*
1561 1562
 * Requeue requests whose mapping to an OSD has changed.  If requests map to
 * no osd, request a new map.
1563
 *
1564
 * Caller should hold map_sem for read.
1565
 */
1566
static void kick_requests(struct ceph_osd_client *osdc, int force_resend)
1567
{
1568
	struct ceph_osd_request *req, *nreq;
1569 1570 1571
	struct rb_node *p;
	int needmap = 0;
	int err;
1572

1573
	dout("kick_requests %s\n", force_resend ? " (force resend)" : "");
1574
	mutex_lock(&osdc->request_mutex);
1575
	for (p = rb_first(&osdc->requests); p; ) {
1576
		req = rb_entry(p, struct ceph_osd_request, r_node);
1577
		p = rb_next(p);
1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595

		/*
		 * For linger requests that have not yet been
		 * registered, move them to the linger list; they'll
		 * be sent to the osd in the loop below.  Unregister
		 * the request before re-registering it as a linger
		 * request to ensure the __map_request() below
		 * will decide it needs to be sent.
		 */
		if (req->r_linger && list_empty(&req->r_linger_item)) {
			dout("%p tid %llu restart on osd%d\n",
			     req, req->r_tid,
			     req->r_osd ? req->r_osd->o_osd : -1);
			__unregister_request(osdc, req);
			__register_linger_request(osdc, req);
			continue;
		}

1596
		err = __map_request(osdc, req, force_resend);
1597 1598 1599 1600 1601 1602
		if (err < 0)
			continue;  /* error */
		if (req->r_osd == NULL) {
			dout("%p tid %llu maps to no osd\n", req, req->r_tid);
			needmap++;  /* request a newer map */
		} else if (err > 0) {
1603 1604 1605 1606
			if (!req->r_linger) {
				dout("%p tid %llu requeued on osd%d\n", req,
				     req->r_tid,
				     req->r_osd ? req->r_osd->o_osd : -1);
1607
				req->r_flags |= CEPH_OSD_FLAG_RETRY;
1608 1609
			}
		}
1610 1611 1612 1613 1614 1615
	}

	list_for_each_entry_safe(req, nreq, &osdc->req_linger,
				 r_linger_item) {
		dout("linger req=%p req->r_osd=%p\n", req, req->r_osd);

1616
		err = __map_request(osdc, req, force_resend);
1617
		dout("__map_request returned %d\n", err);
1618 1619 1620 1621 1622 1623 1624 1625
		if (err == 0)
			continue;  /* no change and no osd was specified */
		if (err < 0)
			continue;  /* hrm! */
		if (req->r_osd == NULL) {
			dout("tid %llu maps to no valid osd\n", req->r_tid);
			needmap++;  /* request a newer map */
			continue;
1626
		}
1627 1628 1629 1630

		dout("kicking lingering %p tid %llu osd%d\n", req, req->r_tid,
		     req->r_osd ? req->r_osd->o_osd : -1);
		__register_request(osdc, req);
1631
		__unregister_linger_request(osdc, req);
1632
	}
S
Sage Weil 已提交
1633 1634 1635 1636 1637 1638
	mutex_unlock(&osdc->request_mutex);

	if (needmap) {
		dout("%d requests for down osds, need new map\n", needmap);
		ceph_monc_request_next_osdmap(&osdc->client->monc);
	}
1639
	reset_changed_osds(osdc);
1640
}
1641 1642


S
Sage Weil 已提交
1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665
/*
 * Process updated osd map.
 *
 * The message contains any number of incremental and full maps, normally
 * indicating some sort of topology change in the cluster.  Kick requests
 * off to different OSDs as needed.
 */
void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg)
{
	void *p, *end, *next;
	u32 nr_maps, maplen;
	u32 epoch;
	struct ceph_osdmap *newmap = NULL, *oldmap;
	int err;
	struct ceph_fsid fsid;

	dout("handle_map have %u\n", osdc->osdmap ? osdc->osdmap->epoch : 0);
	p = msg->front.iov_base;
	end = p + msg->front.iov_len;

	/* verify fsid */
	ceph_decode_need(&p, end, sizeof(fsid), bad);
	ceph_decode_copy(&p, &fsid, sizeof(fsid));
1666 1667
	if (ceph_check_fsid(osdc->client, &fsid) < 0)
		return;
S
Sage Weil 已提交
1668 1669 1670 1671 1672 1673 1674 1675

	down_write(&osdc->map_sem);

	/* incremental maps */
	ceph_decode_32_safe(&p, end, nr_maps, bad);
	dout(" %d inc maps\n", nr_maps);
	while (nr_maps > 0) {
		ceph_decode_need(&p, end, 2*sizeof(u32), bad);
1676 1677
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1678 1679 1680 1681 1682 1683 1684
		ceph_decode_need(&p, end, maplen, bad);
		next = p + maplen;
		if (osdc->osdmap && osdc->osdmap->epoch+1 == epoch) {
			dout("applying incremental map %u len %d\n",
			     epoch, maplen);
			newmap = osdmap_apply_incremental(&p, next,
							  osdc->osdmap,
1685
							  &osdc->client->msgr);
S
Sage Weil 已提交
1686 1687 1688 1689
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
1690
			BUG_ON(!newmap);
S
Sage Weil 已提交
1691 1692 1693 1694
			if (newmap != osdc->osdmap) {
				ceph_osdmap_destroy(osdc->osdmap);
				osdc->osdmap = newmap;
			}
1695
			kick_requests(osdc, 0);
S
Sage Weil 已提交
1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710
		} else {
			dout("ignoring incremental map %u len %d\n",
			     epoch, maplen);
		}
		p = next;
		nr_maps--;
	}
	if (newmap)
		goto done;

	/* full maps */
	ceph_decode_32_safe(&p, end, nr_maps, bad);
	dout(" %d full maps\n", nr_maps);
	while (nr_maps) {
		ceph_decode_need(&p, end, 2*sizeof(u32), bad);
1711 1712
		epoch = ceph_decode_32(&p);
		maplen = ceph_decode_32(&p);
S
Sage Weil 已提交
1713 1714 1715 1716 1717 1718 1719 1720 1721
		ceph_decode_need(&p, end, maplen, bad);
		if (nr_maps > 1) {
			dout("skipping non-latest full map %u len %d\n",
			     epoch, maplen);
		} else if (osdc->osdmap && osdc->osdmap->epoch >= epoch) {
			dout("skipping full map %u len %d, "
			     "older than our %u\n", epoch, maplen,
			     osdc->osdmap->epoch);
		} else {
1722 1723
			int skipped_map = 0;

S
Sage Weil 已提交
1724 1725 1726 1727 1728 1729
			dout("taking full map %u len %d\n", epoch, maplen);
			newmap = osdmap_decode(&p, p+maplen);
			if (IS_ERR(newmap)) {
				err = PTR_ERR(newmap);
				goto bad;
			}
1730
			BUG_ON(!newmap);
S
Sage Weil 已提交
1731 1732
			oldmap = osdc->osdmap;
			osdc->osdmap = newmap;
1733 1734 1735
			if (oldmap) {
				if (oldmap->epoch + 1 < newmap->epoch)
					skipped_map = 1;
S
Sage Weil 已提交
1736
				ceph_osdmap_destroy(oldmap);
1737 1738
			}
			kick_requests(osdc, skipped_map);
S
Sage Weil 已提交
1739 1740 1741 1742 1743 1744 1745 1746
		}
		p += maplen;
		nr_maps--;
	}

done:
	downgrade_write(&osdc->map_sem);
	ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch);
1747 1748 1749 1750 1751 1752 1753 1754 1755

	/*
	 * subscribe to subsequent osdmap updates if full to ensure
	 * we find out when we are no longer full and stop returning
	 * ENOSPC.
	 */
	if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL))
		ceph_monc_request_next_osdmap(&osdc->client->monc);

A
Alex Elder 已提交
1756 1757 1758
	mutex_lock(&osdc->request_mutex);
	__send_queued(osdc);
	mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
1759
	up_read(&osdc->map_sem);
1760
	wake_up_all(&osdc->client->auth_wq);
S
Sage Weil 已提交
1761 1762 1763 1764
	return;

bad:
	pr_err("osdc handle_map corrupt msg\n");
1765
	ceph_msg_dump(msg);
S
Sage Weil 已提交
1766 1767 1768 1769
	up_write(&osdc->map_sem);
	return;
}

1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851
/*
 * watch/notify callback event infrastructure
 *
 * These callbacks are used both for watch and notify operations.
 */
static void __release_event(struct kref *kref)
{
	struct ceph_osd_event *event =
		container_of(kref, struct ceph_osd_event, kref);

	dout("__release_event %p\n", event);
	kfree(event);
}

static void get_event(struct ceph_osd_event *event)
{
	kref_get(&event->kref);
}

void ceph_osdc_put_event(struct ceph_osd_event *event)
{
	kref_put(&event->kref, __release_event);
}
EXPORT_SYMBOL(ceph_osdc_put_event);

static void __insert_event(struct ceph_osd_client *osdc,
			     struct ceph_osd_event *new)
{
	struct rb_node **p = &osdc->event_tree.rb_node;
	struct rb_node *parent = NULL;
	struct ceph_osd_event *event = NULL;

	while (*p) {
		parent = *p;
		event = rb_entry(parent, struct ceph_osd_event, node);
		if (new->cookie < event->cookie)
			p = &(*p)->rb_left;
		else if (new->cookie > event->cookie)
			p = &(*p)->rb_right;
		else
			BUG();
	}

	rb_link_node(&new->node, parent, p);
	rb_insert_color(&new->node, &osdc->event_tree);
}

static struct ceph_osd_event *__find_event(struct ceph_osd_client *osdc,
					        u64 cookie)
{
	struct rb_node **p = &osdc->event_tree.rb_node;
	struct rb_node *parent = NULL;
	struct ceph_osd_event *event = NULL;

	while (*p) {
		parent = *p;
		event = rb_entry(parent, struct ceph_osd_event, node);
		if (cookie < event->cookie)
			p = &(*p)->rb_left;
		else if (cookie > event->cookie)
			p = &(*p)->rb_right;
		else
			return event;
	}
	return NULL;
}

static void __remove_event(struct ceph_osd_event *event)
{
	struct ceph_osd_client *osdc = event->osdc;

	if (!RB_EMPTY_NODE(&event->node)) {
		dout("__remove_event removed %p\n", event);
		rb_erase(&event->node, &osdc->event_tree);
		ceph_osdc_put_event(event);
	} else {
		dout("__remove_event didn't remove %p\n", event);
	}
}

int ceph_osdc_create_event(struct ceph_osd_client *osdc,
			   void (*event_cb)(u64, u64, u8, void *),
1852
			   void *data, struct ceph_osd_event **pevent)
1853 1854 1855 1856 1857 1858 1859 1860 1861
{
	struct ceph_osd_event *event;

	event = kmalloc(sizeof(*event), GFP_NOIO);
	if (!event)
		return -ENOMEM;

	dout("create_event %p\n", event);
	event->cb = event_cb;
1862
	event->one_shot = 0;
1863 1864 1865
	event->data = data;
	event->osdc = osdc;
	INIT_LIST_HEAD(&event->osd_node);
1866
	RB_CLEAR_NODE(&event->node);
1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912
	kref_init(&event->kref);   /* one ref for us */
	kref_get(&event->kref);    /* one ref for the caller */

	spin_lock(&osdc->event_lock);
	event->cookie = ++osdc->event_count;
	__insert_event(osdc, event);
	spin_unlock(&osdc->event_lock);

	*pevent = event;
	return 0;
}
EXPORT_SYMBOL(ceph_osdc_create_event);

void ceph_osdc_cancel_event(struct ceph_osd_event *event)
{
	struct ceph_osd_client *osdc = event->osdc;

	dout("cancel_event %p\n", event);
	spin_lock(&osdc->event_lock);
	__remove_event(event);
	spin_unlock(&osdc->event_lock);
	ceph_osdc_put_event(event); /* caller's */
}
EXPORT_SYMBOL(ceph_osdc_cancel_event);


static void do_event_work(struct work_struct *work)
{
	struct ceph_osd_event_work *event_work =
		container_of(work, struct ceph_osd_event_work, work);
	struct ceph_osd_event *event = event_work->event;
	u64 ver = event_work->ver;
	u64 notify_id = event_work->notify_id;
	u8 opcode = event_work->opcode;

	dout("do_event_work completing %p\n", event);
	event->cb(ver, notify_id, opcode, event->data);
	dout("do_event_work completed %p\n", event);
	ceph_osdc_put_event(event);
	kfree(event_work);
}


/*
 * Process osd watch notifications
 */
1913 1914
static void handle_watch_notify(struct ceph_osd_client *osdc,
				struct ceph_msg *msg)
1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934
{
	void *p, *end;
	u8 proto_ver;
	u64 cookie, ver, notify_id;
	u8 opcode;
	struct ceph_osd_event *event;
	struct ceph_osd_event_work *event_work;

	p = msg->front.iov_base;
	end = p + msg->front.iov_len;

	ceph_decode_8_safe(&p, end, proto_ver, bad);
	ceph_decode_8_safe(&p, end, opcode, bad);
	ceph_decode_64_safe(&p, end, cookie, bad);
	ceph_decode_64_safe(&p, end, ver, bad);
	ceph_decode_64_safe(&p, end, notify_id, bad);

	spin_lock(&osdc->event_lock);
	event = __find_event(osdc, cookie);
	if (event) {
1935
		BUG_ON(event->one_shot);
1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946
		get_event(event);
	}
	spin_unlock(&osdc->event_lock);
	dout("handle_watch_notify cookie %lld ver %lld event %p\n",
	     cookie, ver, event);
	if (event) {
		event_work = kmalloc(sizeof(*event_work), GFP_NOIO);
		if (!event_work) {
			dout("ERROR: could not allocate event_work\n");
			goto done_err;
		}
1947
		INIT_WORK(&event_work->work, do_event_work);
1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968
		event_work->event = event;
		event_work->ver = ver;
		event_work->notify_id = notify_id;
		event_work->opcode = opcode;
		if (!queue_work(osdc->notify_wq, &event_work->work)) {
			dout("WARNING: failed to queue notify event work\n");
			goto done_err;
		}
	}

	return;

done_err:
	ceph_osdc_put_event(event);
	return;

bad:
	pr_err("osdc handle_watch_notify corrupt msg\n");
	return;
}

1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066
/*
 * build new request AND message
 *
 */
void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off,
				struct ceph_snap_context *snapc, u64 snap_id,
				struct timespec *mtime)
{
	struct ceph_msg *msg = req->r_request;
	void *p;
	size_t msg_size;
	int flags = req->r_flags;
	u64 data_len;
	unsigned int i;

	req->r_snapid = snap_id;
	req->r_snapc = ceph_get_snap_context(snapc);

	/* encode request */
	msg->hdr.version = cpu_to_le16(4);

	p = msg->front.iov_base;
	ceph_encode_32(&p, 1);   /* client_inc  is always 1 */
	req->r_request_osdmap_epoch = p;
	p += 4;
	req->r_request_flags = p;
	p += 4;
	if (req->r_flags & CEPH_OSD_FLAG_WRITE)
		ceph_encode_timespec(p, mtime);
	p += sizeof(struct ceph_timespec);
	req->r_request_reassert_version = p;
	p += sizeof(struct ceph_eversion); /* will get filled in */

	/* oloc */
	ceph_encode_8(&p, 4);
	ceph_encode_8(&p, 4);
	ceph_encode_32(&p, 8 + 4 + 4);
	req->r_request_pool = p;
	p += 8;
	ceph_encode_32(&p, -1);  /* preferred */
	ceph_encode_32(&p, 0);   /* key len */

	ceph_encode_8(&p, 1);
	req->r_request_pgid = p;
	p += 8 + 4;
	ceph_encode_32(&p, -1);  /* preferred */

	/* oid */
	ceph_encode_32(&p, req->r_oid_len);
	memcpy(p, req->r_oid, req->r_oid_len);
	dout("oid '%.*s' len %d\n", req->r_oid_len, req->r_oid, req->r_oid_len);
	p += req->r_oid_len;

	/* ops--can imply data */
	ceph_encode_16(&p, (u16)req->r_num_ops);
	data_len = 0;
	for (i = 0; i < req->r_num_ops; i++) {
		data_len += osd_req_encode_op(req, p, i);
		p += sizeof(struct ceph_osd_op);
	}

	/* snaps */
	ceph_encode_64(&p, req->r_snapid);
	ceph_encode_64(&p, req->r_snapc ? req->r_snapc->seq : 0);
	ceph_encode_32(&p, req->r_snapc ? req->r_snapc->num_snaps : 0);
	if (req->r_snapc) {
		for (i = 0; i < snapc->num_snaps; i++) {
			ceph_encode_64(&p, req->r_snapc->snaps[i]);
		}
	}

	req->r_request_attempts = p;
	p += 4;

	/* data */
	if (flags & CEPH_OSD_FLAG_WRITE) {
		u16 data_off;

		/*
		 * The header "data_off" is a hint to the receiver
		 * allowing it to align received data into its
		 * buffers such that there's no need to re-copy
		 * it before writing it to disk (direct I/O).
		 */
		data_off = (u16) (off & 0xffff);
		req->r_request->hdr.data_off = cpu_to_le16(data_off);
	}
	req->r_request->hdr.data_len = cpu_to_le32(data_len);

	BUG_ON(p > msg->front.iov_base + msg->front.iov_len);
	msg_size = p - msg->front.iov_base;
	msg->front.iov_len = msg_size;
	msg->hdr.front_len = cpu_to_le32(msg_size);

	dout("build_request msg_size was %d\n", (int)msg_size);
}
EXPORT_SYMBOL(ceph_osdc_build_request);

2067 2068 2069 2070 2071 2072 2073 2074 2075
/*
 * Register request, send initial attempt.
 */
int ceph_osdc_start_request(struct ceph_osd_client *osdc,
			    struct ceph_osd_request *req,
			    bool nofail)
{
	int rc = 0;

S
Sage Weil 已提交
2076 2077
	down_read(&osdc->map_sem);
	mutex_lock(&osdc->request_mutex);
2078
	__register_request(osdc, req);
2079 2080 2081 2082 2083 2084 2085
	WARN_ON(req->r_sent);
	rc = __map_request(osdc, req, 0);
	if (rc < 0) {
		if (nofail) {
			dout("osdc_start_request failed map, "
				" will retry %lld\n", req->r_tid);
			rc = 0;
S
Sage Weil 已提交
2086
		}
2087
		goto out_unlock;
S
Sage Weil 已提交
2088
	}
2089 2090 2091 2092
	if (req->r_osd == NULL) {
		dout("send_request %p no up osds in pg\n", req);
		ceph_monc_request_next_osdmap(&osdc->client->monc);
	} else {
2093
		__send_queued(osdc);
2094 2095
	}
	rc = 0;
2096
out_unlock:
S
Sage Weil 已提交
2097 2098 2099 2100
	mutex_unlock(&osdc->request_mutex);
	up_read(&osdc->map_sem);
	return rc;
}
2101
EXPORT_SYMBOL(ceph_osdc_start_request);
S
Sage Weil 已提交
2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114

/*
 * wait for a request to complete
 */
int ceph_osdc_wait_request(struct ceph_osd_client *osdc,
			   struct ceph_osd_request *req)
{
	int rc;

	rc = wait_for_completion_interruptible(&req->r_completion);
	if (rc < 0) {
		mutex_lock(&osdc->request_mutex);
		__cancel_request(req);
2115
		__unregister_request(osdc, req);
S
Sage Weil 已提交
2116
		mutex_unlock(&osdc->request_mutex);
S
Sage Weil 已提交
2117
		complete_request(req);
2118
		dout("wait_request tid %llu canceled/timed out\n", req->r_tid);
S
Sage Weil 已提交
2119 2120 2121 2122 2123 2124
		return rc;
	}

	dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result);
	return req->r_result;
}
2125
EXPORT_SYMBOL(ceph_osdc_wait_request);
S
Sage Weil 已提交
2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158

/*
 * sync - wait for all in-flight requests to flush.  avoid starvation.
 */
void ceph_osdc_sync(struct ceph_osd_client *osdc)
{
	struct ceph_osd_request *req;
	u64 last_tid, next_tid = 0;

	mutex_lock(&osdc->request_mutex);
	last_tid = osdc->last_tid;
	while (1) {
		req = __lookup_request_ge(osdc, next_tid);
		if (!req)
			break;
		if (req->r_tid > last_tid)
			break;

		next_tid = req->r_tid + 1;
		if ((req->r_flags & CEPH_OSD_FLAG_WRITE) == 0)
			continue;

		ceph_osdc_get_request(req);
		mutex_unlock(&osdc->request_mutex);
		dout("sync waiting on tid %llu (last is %llu)\n",
		     req->r_tid, last_tid);
		wait_for_completion(&req->r_safe_completion);
		mutex_lock(&osdc->request_mutex);
		ceph_osdc_put_request(req);
	}
	mutex_unlock(&osdc->request_mutex);
	dout("sync done (thru tid %llu)\n", last_tid);
}
2159
EXPORT_SYMBOL(ceph_osdc_sync);
S
Sage Weil 已提交
2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176

/*
 * init, shutdown
 */
int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client)
{
	int err;

	dout("init\n");
	osdc->client = client;
	osdc->osdmap = NULL;
	init_rwsem(&osdc->map_sem);
	init_completion(&osdc->map_waiters);
	osdc->last_requested_map = 0;
	mutex_init(&osdc->request_mutex);
	osdc->last_tid = 0;
	osdc->osds = RB_ROOT;
2177
	INIT_LIST_HEAD(&osdc->osd_lru);
S
Sage Weil 已提交
2178
	osdc->requests = RB_ROOT;
2179
	INIT_LIST_HEAD(&osdc->req_lru);
2180 2181
	INIT_LIST_HEAD(&osdc->req_unsent);
	INIT_LIST_HEAD(&osdc->req_notarget);
2182
	INIT_LIST_HEAD(&osdc->req_linger);
S
Sage Weil 已提交
2183 2184
	osdc->num_requests = 0;
	INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout);
2185
	INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout);
2186 2187 2188
	spin_lock_init(&osdc->event_lock);
	osdc->event_tree = RB_ROOT;
	osdc->event_count = 0;
2189 2190

	schedule_delayed_work(&osdc->osds_timeout_work,
2191
	   round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ));
S
Sage Weil 已提交
2192

2193
	err = -ENOMEM;
S
Sage Weil 已提交
2194 2195 2196
	osdc->req_mempool = mempool_create_kmalloc_pool(10,
					sizeof(struct ceph_osd_request));
	if (!osdc->req_mempool)
2197
		goto out;
S
Sage Weil 已提交
2198

2199 2200
	err = ceph_msgpool_init(&osdc->msgpool_op, CEPH_MSG_OSD_OP,
				OSD_OP_FRONT_LEN, 10, true,
2201
				"osd_op");
S
Sage Weil 已提交
2202
	if (err < 0)
2203
		goto out_mempool;
2204
	err = ceph_msgpool_init(&osdc->msgpool_op_reply, CEPH_MSG_OSD_OPREPLY,
2205 2206
				OSD_OPREPLY_FRONT_LEN, 10, true,
				"osd_op_reply");
S
Sage Weil 已提交
2207 2208
	if (err < 0)
		goto out_msgpool;
2209 2210 2211 2212 2213 2214 2215

	osdc->notify_wq = create_singlethread_workqueue("ceph-watch-notify");
	if (IS_ERR(osdc->notify_wq)) {
		err = PTR_ERR(osdc->notify_wq);
		osdc->notify_wq = NULL;
		goto out_msgpool;
	}
S
Sage Weil 已提交
2216
	return 0;
2217

S
Sage Weil 已提交
2218 2219
out_msgpool:
	ceph_msgpool_destroy(&osdc->msgpool_op);
2220 2221 2222 2223
out_mempool:
	mempool_destroy(osdc->req_mempool);
out:
	return err;
S
Sage Weil 已提交
2224 2225 2226 2227
}

void ceph_osdc_stop(struct ceph_osd_client *osdc)
{
2228 2229
	flush_workqueue(osdc->notify_wq);
	destroy_workqueue(osdc->notify_wq);
S
Sage Weil 已提交
2230
	cancel_delayed_work_sync(&osdc->timeout_work);
2231
	cancel_delayed_work_sync(&osdc->osds_timeout_work);
S
Sage Weil 已提交
2232 2233 2234 2235
	if (osdc->osdmap) {
		ceph_osdmap_destroy(osdc->osdmap);
		osdc->osdmap = NULL;
	}
2236
	remove_all_osds(osdc);
S
Sage Weil 已提交
2237 2238
	mempool_destroy(osdc->req_mempool);
	ceph_msgpool_destroy(&osdc->msgpool_op);
S
Sage Weil 已提交
2239
	ceph_msgpool_destroy(&osdc->msgpool_op_reply);
S
Sage Weil 已提交
2240 2241 2242 2243 2244 2245 2246 2247 2248 2249
}

/*
 * Read some contiguous pages.  If we cross a stripe boundary, shorten
 * *plen.  Return number of bytes read, or error.
 */
int ceph_osdc_readpages(struct ceph_osd_client *osdc,
			struct ceph_vino vino, struct ceph_file_layout *layout,
			u64 off, u64 *plen,
			u32 truncate_seq, u64 truncate_size,
2250
			struct page **pages, int num_pages, int page_align)
S
Sage Weil 已提交
2251 2252 2253 2254 2255 2256
{
	struct ceph_osd_request *req;
	int rc = 0;

	dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino,
	     vino.snap, off, *plen);
2257
	req = ceph_osdc_new_request(osdc, layout, vino, off, plen, 1,
S
Sage Weil 已提交
2258
				    CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
2259
				    NULL, truncate_seq, truncate_size,
2260
				    false);
2261 2262
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2263 2264

	/* it may be a short read due to an object boundary */
2265

2266
	osd_req_op_extent_osd_data_pages(req, 0,
2267
				pages, *plen, page_align, false, false);
S
Sage Weil 已提交
2268

2269
	dout("readpages  final extent is %llu~%llu (%llu bytes align %d)\n",
2270
	     off, *plen, *plen, page_align);
S
Sage Weil 已提交
2271

2272
	ceph_osdc_build_request(req, off, NULL, vino.snap, NULL);
2273

S
Sage Weil 已提交
2274 2275 2276 2277 2278 2279 2280 2281
	rc = ceph_osdc_start_request(osdc, req, false);
	if (!rc)
		rc = ceph_osdc_wait_request(osdc, req);

	ceph_osdc_put_request(req);
	dout("readpages result %d\n", rc);
	return rc;
}
2282
EXPORT_SYMBOL(ceph_osdc_readpages);
S
Sage Weil 已提交
2283 2284 2285 2286 2287 2288 2289 2290 2291 2292

/*
 * do a synchronous write on N pages
 */
int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
			 struct ceph_file_layout *layout,
			 struct ceph_snap_context *snapc,
			 u64 off, u64 len,
			 u32 truncate_seq, u64 truncate_size,
			 struct timespec *mtime,
2293
			 struct page **pages, int num_pages)
S
Sage Weil 已提交
2294 2295 2296
{
	struct ceph_osd_request *req;
	int rc = 0;
2297
	int page_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
2298

2299
	BUG_ON(vino.snap != CEPH_NOSNAP);	/* snapshots aren't writeable */
2300
	req = ceph_osdc_new_request(osdc, layout, vino, off, &len, 1,
S
Sage Weil 已提交
2301
				    CEPH_OSD_OP_WRITE,
2302
				    CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
2303
				    snapc, truncate_seq, truncate_size,
2304
				    true);
2305 2306
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
2307 2308

	/* it may be a short write due to an object boundary */
2309
	osd_req_op_extent_osd_data_pages(req, 0, pages, len, page_align,
2310 2311
				false, false);
	dout("writepages %llu~%llu (%llu bytes)\n", off, len, len);
S
Sage Weil 已提交
2312

2313
	ceph_osdc_build_request(req, off, snapc, CEPH_NOSNAP, mtime);
2314

2315
	rc = ceph_osdc_start_request(osdc, req, true);
S
Sage Weil 已提交
2316 2317 2318 2319 2320 2321 2322 2323 2324
	if (!rc)
		rc = ceph_osdc_wait_request(osdc, req);

	ceph_osdc_put_request(req);
	if (rc == 0)
		rc = len;
	dout("writepages result %d\n", rc);
	return rc;
}
2325
EXPORT_SYMBOL(ceph_osdc_writepages);
S
Sage Weil 已提交
2326 2327 2328 2329 2330 2331 2332

/*
 * handle incoming message
 */
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
{
	struct ceph_osd *osd = con->private;
2333
	struct ceph_osd_client *osdc;
S
Sage Weil 已提交
2334 2335 2336
	int type = le16_to_cpu(msg->hdr.type);

	if (!osd)
S
Sage Weil 已提交
2337
		goto out;
2338
	osdc = osd->o_osdc;
S
Sage Weil 已提交
2339 2340 2341 2342 2343 2344

	switch (type) {
	case CEPH_MSG_OSD_MAP:
		ceph_osdc_handle_map(osdc, msg);
		break;
	case CEPH_MSG_OSD_OPREPLY:
2345
		handle_reply(osdc, msg, con);
S
Sage Weil 已提交
2346
		break;
2347 2348 2349
	case CEPH_MSG_WATCH_NOTIFY:
		handle_watch_notify(osdc, msg);
		break;
S
Sage Weil 已提交
2350 2351 2352 2353 2354

	default:
		pr_err("received unknown message type %d %s\n", type,
		       ceph_msg_type_name(type));
	}
S
Sage Weil 已提交
2355
out:
S
Sage Weil 已提交
2356 2357 2358
	ceph_msg_put(msg);
}

2359
/*
2360 2361
 * lookup and return message for incoming reply.  set up reply message
 * pages.
2362 2363
 */
static struct ceph_msg *get_reply(struct ceph_connection *con,
2364 2365
				  struct ceph_msg_header *hdr,
				  int *skip)
S
Sage Weil 已提交
2366 2367 2368
{
	struct ceph_osd *osd = con->private;
	struct ceph_osd_client *osdc = osd->o_osdc;
2369
	struct ceph_msg *m;
2370
	struct ceph_osd_request *req;
2371 2372
	int front = le32_to_cpu(hdr->front_len);
	int data_len = le32_to_cpu(hdr->data_len);
2373
	u64 tid;
S
Sage Weil 已提交
2374

2375 2376 2377 2378 2379 2380
	tid = le64_to_cpu(hdr->tid);
	mutex_lock(&osdc->request_mutex);
	req = __lookup_request(osdc, tid);
	if (!req) {
		*skip = 1;
		m = NULL;
2381 2382
		dout("get_reply unknown tid %llu from osd%d\n", tid,
		     osd->o_osd);
2383 2384
		goto out;
	}
S
Sage Weil 已提交
2385

2386
	if (req->r_reply->con)
2387
		dout("%s revoking msg %p from old con %p\n", __func__,
2388 2389
		     req->r_reply, req->r_reply->con);
	ceph_msg_revoke_incoming(req->r_reply);
2390

S
Sage Weil 已提交
2391 2392 2393
	if (front > req->r_reply->front.iov_len) {
		pr_warning("get_reply front %d > preallocated %d\n",
			   front, (int)req->r_reply->front.iov_len);
2394
		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS, false);
2395
		if (!m)
S
Sage Weil 已提交
2396 2397 2398 2399 2400 2401
			goto out;
		ceph_msg_put(req->r_reply);
		req->r_reply = m;
	}
	m = ceph_msg_get(req->r_reply);

2402
	if (data_len > 0) {
2403
		struct ceph_osd_data *osd_data;
2404

2405 2406 2407 2408 2409
		/*
		 * XXX This is assuming there is only one op containing
		 * XXX page data.  Probably OK for reads, but this
		 * XXX ought to be done more generally.
		 */
2410
		osd_data = osd_req_op_extent_osd_data(req, 0);
2411 2412
		if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGES) {
			if (osd_data->pages &&
2413
				unlikely(osd_data->length < data_len)) {
2414

2415 2416 2417
				pr_warning("tid %lld reply has %d bytes "
					"we had only %llu bytes ready\n",
					tid, data_len, osd_data->length);
2418 2419 2420 2421 2422 2423
				*skip = 1;
				ceph_msg_put(m);
				m = NULL;
				goto out;
			}
		}
2424
	}
2425
	*skip = 0;
S
Sage Weil 已提交
2426
	dout("get_reply tid %lld %p\n", tid, m);
2427 2428 2429

out:
	mutex_unlock(&osdc->request_mutex);
2430
	return m;
2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441

}

static struct ceph_msg *alloc_msg(struct ceph_connection *con,
				  struct ceph_msg_header *hdr,
				  int *skip)
{
	struct ceph_osd *osd = con->private;
	int type = le16_to_cpu(hdr->type);
	int front = le32_to_cpu(hdr->front_len);

A
Alex Elder 已提交
2442
	*skip = 0;
2443 2444
	switch (type) {
	case CEPH_MSG_OSD_MAP:
2445
	case CEPH_MSG_WATCH_NOTIFY:
2446
		return ceph_msg_new(type, front, GFP_NOFS, false);
2447 2448 2449 2450 2451 2452 2453 2454
	case CEPH_MSG_OSD_OPREPLY:
		return get_reply(con, hdr, skip);
	default:
		pr_info("alloc_msg unexpected msg type %d from osd%d\n", type,
			osd->o_osd);
		*skip = 1;
		return NULL;
	}
S
Sage Weil 已提交
2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473
}

/*
 * Wrappers to refcount containing ceph_osd struct
 */
static struct ceph_connection *get_osd_con(struct ceph_connection *con)
{
	struct ceph_osd *osd = con->private;
	if (get_osd(osd))
		return con;
	return NULL;
}

static void put_osd_con(struct ceph_connection *con)
{
	struct ceph_osd *osd = con->private;
	put_osd(osd);
}

2474 2475 2476
/*
 * authentication
 */
2477 2478 2479 2480 2481
/*
 * Note: returned pointer is the address of a structure that's
 * managed separately.  Caller must *not* attempt to free it.
 */
static struct ceph_auth_handshake *get_authorizer(struct ceph_connection *con,
2482
					int *proto, int force_new)
2483 2484 2485 2486
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;
2487
	struct ceph_auth_handshake *auth = &o->o_auth;
2488

2489
	if (force_new && auth->authorizer) {
2490
		ceph_auth_destroy_authorizer(ac, auth->authorizer);
2491 2492
		auth->authorizer = NULL;
	}
2493 2494 2495
	if (!auth->authorizer) {
		int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
						      auth);
2496
		if (ret)
2497
			return ERR_PTR(ret);
2498 2499
	} else {
		int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_OSD,
2500 2501 2502
						     auth);
		if (ret)
			return ERR_PTR(ret);
2503 2504
	}
	*proto = ac->protocol;
2505

2506
	return auth;
2507 2508 2509 2510 2511 2512 2513 2514 2515
}


static int verify_authorizer_reply(struct ceph_connection *con, int len)
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;

2516
	return ceph_auth_verify_authorizer_reply(ac, o->o_auth.authorizer, len);
2517 2518
}

2519 2520 2521 2522 2523 2524
static int invalidate_authorizer(struct ceph_connection *con)
{
	struct ceph_osd *o = con->private;
	struct ceph_osd_client *osdc = o->o_osdc;
	struct ceph_auth_client *ac = osdc->client->monc.auth;

2525
	ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD);
2526 2527
	return ceph_monc_validate_auth(&osdc->client->monc);
}
2528

2529
static const struct ceph_connection_operations osd_con_ops = {
S
Sage Weil 已提交
2530 2531 2532
	.get = get_osd_con,
	.put = put_osd_con,
	.dispatch = dispatch,
2533 2534
	.get_authorizer = get_authorizer,
	.verify_authorizer_reply = verify_authorizer_reply,
2535
	.invalidate_authorizer = invalidate_authorizer,
S
Sage Weil 已提交
2536
	.alloc_msg = alloc_msg,
2537
	.fault = osd_reset,
S
Sage Weil 已提交
2538
};